The Vision and Media Lab
Image-based Music Generation
This work studies converting image content to music using visual-auditory assoications. Image features, such as contour, color and texture, are extracted and converted to music segments in various methods. Music segments can be integrated to form longer and more interesting music patterns.

To listen to some experimental results, please follow this link.

Xiaoying Wu and Ze-Nian Li. Image-based Music Composition . IEEE International Conference on Multimedia and Expo , Jun 2008. [pdf]